Loss of discrimination on an LV switchboard with a 30 kA fault level
A 1,250 kVA, 11 kV/400 V Dyn11 transformer with 5.5% impedance feeds a main LV switchboard through a 1,600 A air circuit breaker. Outgoing ways are a 630 A moulded case breaker to a riser, a 400 A breaker to a chiller plant room and a 250 A breaker to a tenant distribution board; the largest tenant submain is a 63 A type C MCB protecting 35 mm² copper XLPE singles on a 45 m run. Commissioning gives a prospective fault current of 30.2 kA at the main board. When the contractor tested a short circuit at the tenant board terminals, both the 250 A breaker and the main air circuit breaker tripped, blacking out the building. The air circuit breaker is set to 1.0×In long time, 4×In short time with a 0.1 s delay, and 8×In instantaneous; the contractor says this 'protects the board properly and makes it fast'. The client has withheld acceptance of the switchboard pending your written assessment.
1. Calculate the transformer full-load and prospective short-circuit currents, and explain why the main air circuit breaker tripped despite the 0.1 s short-time delay.
Read the worked answer
Fault current check
- Full-load current: I = 1,250,000 / (√3 × 400) = 1,804 A.
- Prospective short-circuit current at the transformer terminals, neglecting cable impedance: Isc ≈ I / (Z%) = 1,804 / 0.055 = 32.8 kA.
- The measured 30.2 kA at the board is consistent once busbar and cable impedance are included. The ACB's rated short-circuit capacity (Icu) and its short-time withstand (Icw for 0.1 s) must both be at least this value; typical 1,600 A devices are built for 42–50 kA, so check the nameplate and the type-test certificate rather than assuming.
Why the ACB tripped
The instantaneous element is set to 8 × 1,600 A = 12.8 kA. The fault current at the tenant board lies well above that, so the ACB's instantaneous element operates in a few tens of milliseconds and completely bypasses the 0.1 s short-time delay. The intended time discrimination with the 250 A breaker never gets a chance to work. Turning the instantaneous setting down to make the board 'fast' is exactly the wrong direction: speed at the top of the system destroys selectivity.
Correct approach
Switch the instantaneous element off (or set it above the maximum prospective fault current at that outgoing way), keep the short-time delay coordinated with the downstream breaker's trip band, and verify the pair against the manufacturer's published discrimination tables. Zone-selective interlocking is an option if the tables show insufficient time margin at the highest fault level.
2. Assess the 35 mm² submain and the 63 A type C MCB for the fault conditions at the tenant board, and state what changes are needed.
Read the worked answer
Cable thermal withstand (adiabatic check)
Use S = √(I²t) / k, with k = 143 for copper XLPE. If the fault current at the tenant board is about 20 kA and clearance takes 0.1 s, then I²t = 4 × 10⁷ A²s, and the minimum area is √(4 × 10⁷) / 143 ≈ 44 mm². The 35 mm² conductor would be under-protected on that basis.
The conclusion depends entirely on which device clears and how fast. If a current-limiting MCCB clears in about 10 ms with a let-through I²t of, say, 4 × 10⁶ A²s, the required area drops to about 14 mm² and 35 mm² is adequate. So the design must state the clearing device, its actual let-through energy at the prospective fault current, and the resulting check - not assume the upstream 0.1 s delay applies to the cable.
MCB breaking capacity
The fault level at the tenant board is of the order of 15–20 kA, but a 63 A type C MCB typically has a rated breaking capacity of only 6 kA or 10 kA. It cannot be used there unless back-up protection by the 250 A MCCB is verified from the manufacturer's tested combination tables, or unless the prospective fault current is reduced (longer run, larger cable) or an MCCB replaces it.
Actions
Re-run the fault calculation at the tenant board, select the MCB with adequate Icn or documented back-up, confirm the cable's I²t withstand for the actual clearing device, and record all of it in the protection study.
3. List the settings and documentation you would require before recommending acceptance of the switchboard.
Read the worked answer
Acceptance should be conditional on evidence, not assurances:
- A protection discrimination study covering every outgoing way against the incomer, at minimum and maximum fault levels and at the relevant transformer taps, using the device manufacturers' published tables or software.
- Verified ratings: Icu and Icw of the ACB and MCCBs against the calculated 30.2 kA, and type-test certification of the assembly to IEC 61439 with the correct internal segregation and forms of separation.
- Locked and recorded settings: a settings sheet for every device, with the instantaneous element disabled or set above the outgoing fault level, and physical locking or password protection so site staff cannot drift from the study.
- A functional demonstration of discrimination at the highest practical fault point, or at least secondary-injection testing of the settings.
- Labels on the board showing the fault level, the settings and the date of the study, plus the study revision referenced in the O&M manual.
The contractor's statement that the settings 'protect the board' is not wrong in the narrow thermal sense - the board itself is protected - but protection of the board is not the same as selectivity for the installation. Recommend acceptance only against the documented study, and note the residual risk in writing if the client chooses to accept without it.